• Save
Backhaul Requirements for Centralized and  Distributed Cooperation Techniques
Upcoming SlideShare
Loading in...5
×

Like this? Share it with your network

Share
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
  • This Very useful to me to complete my project i also get very good helpful information from below may be it useful to you all So i Like To share it with you,

    www.teletopix.org
    Are you sure you want to
    Your message goes here
No Downloads

Views

Total Views
2,112
On Slideshare
2,112
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
0
Comments
1
Likes
3

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. Backhaul Requirements for Centralized andDistributed Cooperation TechniquesDr. Stefan BrückQualcomm CDMA Technologies 8th July 2010
  • 2. Overview The LTE Architecture The X2 Interface in LTE Release 8 An introduction to the X2AP protocol Categories of Downlink Coordinated Multipoint (CoMP) Two Implementation Examples Centralized Joint Transmission Decentralized Coordinated Scheduling/Beamforming X2AP Protocol Extensions Examples for the considered CoMP schemes*) Bandwidth and Delay Requirement Analysis for Downlink CoMP Conclusions *) S. Brueck, L. Zhao, J. Giese, M. A. Awais, “Centralized Scheduling for Joint-Transmission Coordinated Multi-Point in LTE-Advanced” and J. Giese, M. A. Awais, “Performance Upper Bounds for Coordinated Beam Selection in LTE-Advanced” in Proceedings of the ITG/IEE Workshop on Smart Antennas, Bremen (WSA10), Germany, 23. – 24. February 20102 2
  • 3. The LTE Architecture MME/S-GW MME/S-GW The core network is packet- switched only The E-UTRAN consists of one node only EPC S1 interface between S-GW/MME S1 S1 and eNB S1 S1 S1 Logical many-to-many interface Supports procedures to establish, S1 maintain and release E-UTRAN Radio Access Bearers E-UTRAN Supports transfer of NAS signalling messages between UE and EPC eNB X2 eNB X2 interface between eNBs Logical point-to-point interface X2 X2 Seamless mobility eNB Interference management EPC = Evolved Packet Core Load management3 3
  • 4. X2 Protocol Structure Radio Control Plane User Plane Network Layer User Plane X2-AP PDUs Transport Transport Network Transport Network User Plane Network User Plane Layer Signaling Transport GTP-U SCTP UDP IP Data IP Data link layer Transport Data link layer Physical layer Physical layer Clear separation between radio network and transport network layers The radio network layers defines interaction between eNBs The transport network layer provides services for user plane and signaling transport4 4
  • 5. X2 Application Protocol (X2AP) The X2AP is responsible for providing signaling transport between eNBs Specification: 3GPP TS 36.423 X2AP functions are executed by so called Elementary Procedures Release 8 defines eleven EPs related to different X2AP functions In Release 9 six additional EPs have been added In Release 8/9 limited load management functionality is supported eICIC is currently under specification for Release 10 Function Elementary Procedure(s) Mobility Management a) Handover Preparation b) SN Status Transfer c) UE Context Release d) Handover Cancel Load Management a) Load Indication b) Resource Status Reporting Initiation Release 8 c) Resource Status Reporting Reporting of General Error Situations Error Indication Resetting the X2 Reset Setting up the X2 X2 Setup eNB Configuration Update a) eNB Configuration Update b) Cell Activation Mobility Parameters Management Mobility Settings Change Mobility Robustness Optimisation a) Radio Link Failure Indication Release 9 b) Handover Report Energy Saving a) eNB Configuration Update b) Cell Activation5 5
  • 6. Coordinated Multipoint in 3GPP Network coordination became popular by publications of Bell Labs (Network MIMO) M. Karakayali, G. Foschini and R. Valenzuela, “Network Coordination for spectrally efficient Communications in Cellular Systems”, August 2006, IEEE Wireless Communications Magazine 3GPP considered coordination techniques under the name Coordinated Multipoint (CoMP) for Release 10 3GPP TR 36.814, v9.0.0, March 2010 3GPP Conclusion: CoMP will not be part of the Release 10 specification 3GPP Downlink CoMP terminology Joint Processing Coordinated Multipoint: User data to be transmitted to one terminal is available in multiple sectors of the network. A subclass of joint processing is joint transmission, where the data channel to one terminal is simultaneously transmitted from multiple sectors. Coordinated Scheduling/Beamforming: User data is only available in one sector, the so- called serving cell, but user scheduling and beamforming decisions are made with coordination among the sectors.6 6
  • 7. Example 1: Centralized Joint Transmission A central master cell manages all Master Cell resources of the cooperating cells Transmission Point 1 Each UE reports channel state = Serving Cell = Master Cell information (CSI) to its serving cell The serving cell forwards the CSI toTransmission Point 1 the master over the X2 interface = Serving Cell Transmission Point 2 The master distributes scheduling = Slave Cell decisions to the transmitting cells over X2 Advantage: Allows optimal scheduling since the master knows the CSI of all UEs Only two X2 usages are required Disadvantage CSI needs to be sent over X2 for each active UE7 7
  • 8. Example 2: Decentralized Coordinated Scheduling Goal: Avoidance of interference in the spatial domain Idea: Identify worst interferer and avoid “collision” Prevent interferers from using “most destructive” precoding matrices Interfering cell can serve UEs on “non-destructive” beams Typical Coordination Approach 1) Active UEs send information about “most destructive” interfering precoding matrices to their serving cells RESTRICTION REQUEST message in slide 9 2) The serving cell forwards the received messages to the interfering cells over X2 Option: The message is only forwarded if the requesting user is scheduled in order to reduce the number of X2 messages 3) Optional: Cells negotiate their beams to improve a utility metric REQUEST GRANT/REJECT messages in slide 98 8
  • 9. Example of a Coordination Time Line*) UE 1 Cell 1 Cell 2 UE 2 UE 3 Cell 3 CSI feedback CSI feedback CSI feedback RESTRICTION REQUEST RESTRICTION REQUEST Do individual Do individual Do individual scheduling; scheduling; scheduling; check if result check if result check if result contradicts RR contradicts RR contradicts RR REQUEST GRANT/REJECT REQUEST GRANT/REJECT if REQUEST GRANT received from cell 1 conflicts with RG sent to cell 3: compare coordination gain with cell 3 to gain with cell 1 send GRANT REJECT to cell with lower coordination gain (here cell 1) if gain with cell 3 > gain with cell 1, send GRANT REJECT Data transmission Data transmission Data transmission Uplink Downlink Cell-to-cell *) Based on J. Giese, M. A. Awais, “Performance Upper Bounds for Coordinated Beam Selection in LTE-Advanced”, in Proceedings of the ITG/IEE Workshop on Smart Antennas, Bremen (WSA10), Germany, 23. – 24. February 20109 9
  • 10. X2 Bandwidth Considerations Centralized Joint Transmission CoMP Estimate of Signaling Load per active User (Slave → Master) Information Content Required Bits Two Cell IDs (ECGI) 56 bits ACK/NACK report 2 x 1 bit (dual stream transmission) 9 bits (subband CQI) CQI/PMI/RI reports (assuming implicit periodic reporting as in LTE Rel8) ≤ 11 bits (wideband CQI) (per measured radio link) Σ ≤ 58 bits + 11 bits/measured radio link Decentralized Coordinated Scheduling/Beamforming Estimate of Signaling Load for Restriction REQUEST Information Content Required Bits Two Cell IDs (ECGI) 56 bits UE average throughput 16 bits SINR Increase by precoding matrix 16 bits restriction (silencing) Σ ≤ 86 bits If coordination takes place each TTI = 1ms, one RESTRICTION REQUEST message requires a bandwidth of 86 kbits/s10 10
  • 11. Frequency of X2 Messages The figures show the number of sent and received X2 messages/cell for a coordinated scheduling approach*) With probability of 95%, less or equal than three messages are received per cell The number of received RESTRICTION REQUEST messages scales with the number of scheduled users in this example Receiving three messages/cell requires a bandwidth of 3 * 86 kbits/s = 258 kbits/s A three sector eNB would therefore require a X2 bandwidth of 3 * 258 kbits/s = 774 kbits/s This assumes 10 UEs/cell and RESTRICTION REQUEST messages targeting 900 kHz and sent at for geometries < - 3dB *) Based on J. Giese, M. A. Awais, “Performance Upper Bounds for Coordinated Beam Selection in LTE-Advanced”, in Proceedings of the ITG/IEE Workshop on Smart Antennas, Bremen (WSA10), Germany, 23. – 24. February 201011 11
  • 12. Impact of X2 Delay The figure shows the throughput loss for a centralized non-coherent JT CoMP approach*) The X2 delay increases the time till the CSI/CQI is available in the MAC scheduler Consequently, the CSI information alters and the data rates cannot be adapted well to the channel The figure shows the impact of the one-way X2 delay V = 3 km/h Release 8 CQI delay = 6ms The X2 delay comprises of a) Interface propagation delay b) eNB Tx/Rx processing delay *) S. Brueck, L. Zhao, J. Giese, M. A. Awais, “Centralized Scheduling for Joint-Transmission Coordinated Multi-Point in LTE-Advanced”, in Proceedings of the ITG/IEE Workshop on Smart Antennas, Bremen (WSA10), Germany, 23. – 24. February 201012 12
  • 13. Conclusions Downlink coordination techniques require extensions of the X2 application protocol (3GPP TS 36.423) In case of centralized coordination, the required X2 bandwidth scales with the number of users in RRC_CONNECTED mode This number can be in the order of tens to hundreds The required X2 bandwidth has a large variance Decentralized coordination techniques allow for a scaling with the number of scheduled users This allows reducing the X2 bandwidth significantly The drawback is a loss in performance compared to optimal schemes Coordination techniques are delay sensitive X2 delay comprises of interface propagation delay and eNB processing delay A X2 delay of 5ms caused a loss in spectral efficiency of 20% for the investigated scheme13 13
  • 14. Thank you!! Contact: Dr. Stefan Brück Tel: 0911 54013270 sbrueck@qualcomm.com14 14